The ‘Eyes ++ actor’ what a beauty!
This Isdaora actor is an incredible piece of coding and when you discover what it can do, it takes motion tracking to another level. The ‘Eyes’ actor is great and is basically a cut down version of the ‘Eyes ++’.
The ‘Eyes’ and ‘Eyes ++’ actors are basically a way of motion tracking objects from the physical world. By watching pixels from a video or more commonly a live feed source, the actors (with a little help sometimes from an effects mixer and a freeze) can track the largest and brightest objects in the frame.
Things to consider.
If you are using a camera at torso/head level then if two or more people walk behind each other – the camera will only see one person (or object). In a way – computers are not clever enough to know that there is a person out of shot and that they will reappear. This can cause problems if a specific person is controlling a specific effect or parameter.
Shadows. A motion tracking nightmare. You can avoid this by using thresholds, blocking out all natural light, or using color tracking instead but unfortunately it can be a little tricky. Experimenting and rehearsing is the only real way to avoid this problem.
Lighting. Closely linked to shadows but basically changing a lighting cue in a performance will alter the live feed video input. Since ‘Eyes’ watches for the largest collection of pixels… a fully blown white wash or spot light can cause problems. Again, rehearsing, planning and foreword thinking solves these issues. Camera Position. Consider locating the camera higher up or mounted safely in the rafters/ceiling if possible. This can solve other lighting and shadow problems but can also produce interesting results in terms of depth and width or (or X&Y or Horizontal & Vertical).
Sometimes, and its only on the odd occasion I have done this… consider knocking the camera out of focus. You can do this when you do not need the audience to see the actual video input. This can be useful to make a person or object become a little larger, due to the blurred pixels at the edges. This works like a natural edge feather and is sometimes. Something to consider anyway.
The Actor (…and its blobs!)
The image above is the default and untouched Eyes ++ actor. I will not be going into great depth about this actor, just enough to get you going. The next post will show you more and will follow soon.
I always turn on the ‘monitor’ this is located at the bottom of the actor on the left hand side. This shows us exactly what the actor itself is seeing. Turn this on and you should see it spring to life.
I then changed the Objects input. from 16 to 1. Now, if you have read my other tutorials (which you should do! hehe) then you will know all about thresholds etc, so I am going to skip that part and set mine to about 14. Set yours to an appropriate level to get rid of all the background noise.
You should now get some ‘basic’ data…. but you can’t see anything coming out of the output of the actor, what the hell are all these blobs?! Well calm down and let me explain.
I guess this is the fundamental part. The 16 blobs on the right hand side of the actor relate to the 16 maximum objects that one Eyes++ actor can watch/track. Since this actor would be huge to show all the information for all 16 possible objects a separate individual actor is used. And this is how…
Makes sense? Obviously you can add all 16 if needed but for now I am going to leave it as just one.
Similar to the ‘Eyes’ actor you get a lot of data, horizontal and vertical values, Objects full Height and Width, Velocity (or speed) of the object and a few others.
You can smooth the values out with the ‘smoothing’ input on the Eyes++ actor.
Have a play around with the ‘bounds track’ and ‘median fit’ to explore what they do and don’t forget that you can alt click on the white text and go to ‘Actor input Help’ to read more about the options. For example, the lifespan option:
I hope this post helps all you interactive wizards! Get playing around and see what you can come up with!